This systematic review was conducted under the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [17].
2.1 Eligibility criteria
Studies that met the following inclusion criteria were included in the systematic review and meta-analysis. The inclusion criteria considered in this review include: -
-
Study Population: All populations regardless of their age, occupation, ethnicity, gender, etc.
-
Outcomes The articles aimed to determine the hesitancy and/or acceptance of COVID-19 vaccines that provided a quantitative outcome.
-
Language: Studies written in English.
-
Types of articles: Peer-reviewed full text, original, and published articles.
-
Publication year: Studies published since after the emergency of COVID-19 to the study period (March 2021).
-
Study region / location: Not specified (no limitation).
However, unpublished manuscripts or preprints; editorial papers; reports; short communication; review articles; the article did not provide quantitative results or did not aim to determine acceptance and / or hesitancy of the COVID-19 vaccine.
2.2 Information sources and search strategy
Article searches were performed using key terms of review such as COVID-19, vaccine hesitancy, vaccine acceptance and intention to vaccine, and Medical Subject Headings (MeSH) in combination with Boolean logic operators (“AND”, “OR”, and “NOT”) through various electronic databases from PubMed/Medline, Scopus or Science direct, Web of Science, Embase, CINAHL, and Google scholars. Furthermore, the articles were searched directly from Google. On the other hand, references within eligible papers were screened for additional articles. There was no specific date for the search of the articles for each database. For example, the articles were searched from first February to March 29, 2021 from PubMed, Scopus, Embase, and Google scholars while the search on Web of Science, CINAHL, and Google was made from 15 February to 31 March 2021.
The Embase search strategy was made as follows: (1) COVID 19 ' OR ‘COVID 19’/exp OR Coronavirus OR coronavirus/exp OR ‘2019 nCOV’ OR ‘2019 nCOV’/exp OR ‘severe acute respiratory syndrome coronavirus 2' / exp OR 'SARS-COV-2' OR 'SARS-COV 2/exp; (2) Vaccine OR vaccine/exp OR Vaccination OR Vaccination/exp OR Immunization OR Immunization/exp; (3) acceptance OR acceptance /exp; OR ‘hesitance’ OR ‘hesitance/exp’ OR ‘refusal OR ‘refusal/exp’;(4) 1 AND 2 AND 3.
The PubMed search strategy was performed as follows: (1) 'Coronavirus' [MeSH Terms] OR 'Coronavirus' [All Fields] OR 'Coronavirus' [All Fields] OR 'COVID-19' [All Fields] OR 'SARS-2' [All Fields] OR 'Severe acute respiratory syndrome coronavirus 2' [All Fields] OR '2019 nCOV' [All Fields] OR 'SARS COV-2' [All Fields] OR 'Coronavirus' [All Fields] AND; (2) 'Vaccines' [MeSH Terms] OR 'Vaccine' [All Fields] OR 'Vaccinations' [All Fields] OR 'Vaccines' [All Fields] OR 'Vaccines' [All Fields]; (3) “acceptance” [MeSH Terms] OR “hesitancy” [All Fields] OR “refusal” [All Fields] OR “accepted” [All Fields] OR “willingness to accept” [All Fields]; (4) 1 AND 2 AND 3.
The Scopus and Web of Science search strategy was made using a combination of keywords and Boolean functions: (1) (COVID-19 * OR Coronavirus* OR nCOV * OR SARS COV-2; (2) Vaccines *OR Vaccination *OR vaccinates); (3) (acceptance *OR hesitancy *OR refusal* OR accepted *OR willingness to accept; (4) 1 AND 2 AND 3) or 1 AND (2 OR 3)
The search strategy from Google Scholars and Google was done using keywords such as (Coronavirus, nCOV, SARS COV-2, COVID-19) AND (Vaccines OR Vaccination OR Vaccinates); (3) (acceptance OR hesitancy OR refusal OR accepted OR willingness to accept; (4) 1 AND 2 AND 3.
Articles published since the emergency of COVID-19 to March 2021 (from inception to March 2021) were searched from the included electronic databases according to their own searching strategies. The search was carried out from the first February 2021 to the last March 2021.
2.3 Study Selection
The study selection process was performed using the PRISMA flow chart, indicating the number of articles included in the review and articles excluded from the study with the reasons. Following the search for articles through selected electronic databases, duplicate studies were removed using the ENDNOTE software version X5 (Thomson Reuters, USA). After removing duplicate articles, the authors (DAM, YAA, and YMD) independently selected the articles based on the titles and abstracts applying the inclusion criteria. Furthermore, the full text of the relevant articles was further read in detail and the inclusion criteria independently evaluated by the authors. Any disagreements made with respect to the inclusion of studies were resolved by consensus after discussion. Finally, studies that met the criteria were included in the systematic review and meta-analysis.
2.4 Data Extraction
The data were extracted by the authors (DAM, YAA, and YMD) independently. Predetermined tabular format consisting of study characteristics including publication year, survey period, country where the study was conducted, number of respondents and outcome (COVID-19 vaccine acceptance/hesitancy rate using Microsoft Excel, 2016). Any disagreement made between the authors was resolved through discussion after the same procedures were repeated.
2.5 Data Quality Assessment/Management
The selected articles were subjected to a rigorous independent assessment using a standardized critical assessment tool, Joanna Briggs Institute (JBI) Critical Assessment Tools for prevalence studies [18]. The evaluation tools have the following nine evaluation criteria/ parameters; (1) appropriate sampling frame; (2) proper sampling technique; (3) adequate sample size; (4) description of the study subject and setting description; (5) sufficient data analysis; (6) use of valid methods for identified conditions; (7) valid measurement for all participants; (8) use of appropriate statistical analysis and (9) adequate response rate.
The authors ((DAM, YAA, and YMD)) assessed the quality of the included studies. Based on the items in the above appraisal tool, the articles were classified as high quality (80% and above), moderate (60–80% score), and low quality (< 60% score). Articles with a score ≥ 60% were included in the review, while those with a low quality were excluded from the study (Supplementary Materials I). Finally, the disagreements made among the authors were resolved by discussion and repeating the same procedures.
2.5 Outcome Measures
The term 'vaccine hesitancy' refers to 'delay in acceptance or refusal of vaccines despite the availability of vaccine services' [6, 19, 20]. In this review, for articles that did not provide general acceptance of the vaccine among study participants, the prevalence of vaccine acceptance was calculated based on the response of the participants. The participant responded strongly agree, agree, completely agree, accept all, accept some, accept, and yes to the questions were considered as accepted. Finally, the prevalence was calculated based on the frequency of responses and the total number of respondents. The same principle was applied to the studies that reported the result based on the Likert scale and others (Fig. 1).
2.1 Statistical Procedures and Data Analysis
The pooled prevalence / acceptance rate of the COVID-19 vaccine was performed using Comprehensive Meta-Analysis (CMA) version 3.0 statistical software. Forest plot and random-effects model were used to determine and visualize the pooled acceptance rate of the COVID-19 vaccine (supplementary file II).
The Cochran Q test (Q) and I Squared test (I2 statistics) were used to evaluate the heterogeneity between the included articles. The, heterogeneity was classified into low (I2 index < 25%), medium (I2 index ranging from 25–75%), and high heterogeneity (I2 index > 75%). The random-effects model was used to analyze the data. Furthermore, subgroup analysis was performed based on the year of publication, survey period (when the study was conducted), and study areas. Publication bias between the included studies was evaluated using funnel plots (Fig. 2). A P-value of < 0.05 was considered as evidence of publication bias. Sensitivity analysis was done to determine differences in pooled effects by dropping studies that were found to influence the summary estimates. Furthermore, meta-regression was conducted to determine the bias as the result of high heterogeneity in the current study.